The data mining methods required to build the approach based on the discipline involved are then defined. The data-mining method depends on whether or not dependent variables are present in the analysis. Predictions with dependent variables (labels) are generated through supervised learning, which can be performed by the use of linear regression, generalized linear regression, a proportional hazards model a competitive risk model, decision trees, the random forest algorithm, and support vector machines. In contrast, unsupervised learning involves no labels. The learning model infers some internal data structure. Common unsupervised learning methods include principal component analysis association analysis, and clustering analysis (Wen Tao Wu 2021)